Autonomous Execution
Autonomous Execution
Section titled “Autonomous Execution”Auteryn agents plan and execute multi-step tasks — calling integrations, searching knowledge, running sandbox tools, and streaming results back — without hand-holding each step.
Execution limits: Fast runs timeout at ~11 minutes. Deep runs timeout at ~2 hours per execution. If a run is interrupted, recovery context can inject plan and scratchpad state on the next message.
Execution modes
Section titled “Execution modes”| Mode | Default for | Timeout (approx.) | Best for |
|---|---|---|---|
| Auto | Internal agents | SmartRouter picks Fast or Deep | Let the platform balance speed vs. depth |
| Fast | Customer-facing | ~11 min | Chat, lookups, quick automation |
| Deep | Internal (when selected) | ~2 hours | Research, multi-step coding, complex workflows |
Customer-facing agents always use Fast mode.
SmartRouter (Auto mode)
Section titled “SmartRouter (Auto mode)”When an Internal agent is set to Auto, the SmartRouter classifies each incoming task and routes to Fast or Deep automatically. Override to Fast or Deep in Agent → Settings → Execution Mode if you want consistent behavior.
What happens during a run
Section titled “What happens during a run”- Agent receives your message (Console, API, flow, or channel)
- Retrieves relevant knowledge chunks (if attached)
- Plans steps and calls tools (integrations, sandbox, browser, web search)
- Streams text and tool events back via SSE
- On completion, async jobs may update conversation summary and memory
Recovery context
Section titled “Recovery context”If a Deep run is interrupted (timeout, cancellation, worker restart), the next message can include recovery context: the previous plan, recent tool results, and scratchpad notes. Enable/disable via platform config (ENABLE_RECOVERY_CONTEXT).
This is within-conversation resumption — not multi-day task persistence across unrelated sessions.